197 results about "Row vector" patented technology

The invention provides a vehicle fault root cause diagnosis method, equipment, equipment and a medium, and the method comprises the steps: obtaining a vehicle target fault phenomenon, determining a to-be-detected fault event set based on a preset fault logic relation, constructing an initial scoring matrix, enabling a row vector to correspond to a fault event, enabling a column vector to correspond to the state feature parameters of a plurality of evaluation dimensions, and carrying out the detection of the fault event set; executing the detection task of the highest comprehensive score fault event, obtaining feedback data, updating the initial score matrix parameters based on the feedback data to obtain an updated score matrix, and calculating the comprehensive score of each event based on the updated score matrix again. Iteratively executing the detection task corresponding to the updated highest score event to identify whether the detection event is a fault root cause or not until the fault root cause of the target fault phenomenon is determined; according to the method, through a dynamic priority scheduling mechanism, real-time feedback data is fused in multi-dimensional evaluation, a high-value diagnosis task is executed preferentially, the resource consumption of traditional traversal diagnosis is remarkably reduced, and the fault positioning efficiency is effectively improved.

The invention belongs to the technical field of rapid calculation of a temperature field of a transformer winding, and particularly relates to a method for extracting a transient temperature rise main mode of the transformer winding by using a mixed norm. Comprising the following steps: S1, setting simulation step length and time, and determining a winding temperature rise sampling analysis point; s2, acquiring temperature rise distribution of a winding temperature field from the first sampling moment to the + 1th sampling moment, and constructing matrixes Tk and Tk + 1; s3, calculating D according to dynamic modal decomposition; s4, calculating an infinity-norm and a 2-norm of the matrix row vector, and defining a convex combination of the infinity-norm and the 2-norm as a norm of the matrix row vector; s5, simplifying a calculation formula of Tk; s6, calculating to obtain an instantaneous value of Tk, setting the instantaneous value to be equal to + 1, returning to the step S3, and carrying out winding temperature rise prediction at the next moment; according to the method, the overall and local characteristics of the modes are considered, the accuracy of main mode selection is improved, and the prediction precision of the instantaneous temperature rise of the winding temperature field is improved.

The invention relates to the field of wastewater treatment, in particular to an online analysis method and system for ink-jet printing wastewater treatment data, and the method comprises the steps: obtaining the water quality, process and equipment state parameter data of ink-jet printing wastewater based on each sensor, and constructing an original data matrix through preprocessing; decomposing the matrix, selecting a preset number of principal components for reconstruction, and generating a reconstructed data matrix; a reconstruction error is calculated by analyzing the difference between an original matrix and a reconstruction matrix, a correlation distortion index is analyzed based on row vector similarity, and an objective function is constructed to determine the number of optimal principal components; and clustering is carried out by using reconstruction errors and distances, abnormal data points are screened, and process parameters are regulated and controlled. According to the method, by optimizing the number of principal components and improving clustering analysis, the noise interference problem under complex water quality changes is effectively solved, the anomaly detection precision and the system stability are remarkably improved, and the robustness and the adaptability of an ink-jet printing wastewater treatment system are enhanced.

An operation accelerator, a processing method, and a related device, the operation accelerator including a first memory configured to store an input dataset, a matrix converter configured to perform reading M row vectors from the input dataset, generating a first instruction, and sending the M row vectors and the first instruction to a second memory configured to perform, according to the first instruction, preprocessing on the M row vectors to obtain n row vectors, and storing the n row vectors, where the n row vectors include the M row vectors and (n−M) padding row vectors, the n row vectors are N row vectors in a target matrix, and a storage sequence of the n row vectors in the second memory is consistent with a sequence of the N row vectors in the target matrix.

The invention discloses a large model training method based on second-order matrix optimization, and belongs to the field of deep learning model training technology optimization, and the large model training method based on second-order matrix optimization comprises the following steps: S1, decomposing a second-order matrix into row and column vectors, and carrying out sliding average and distributed block reduction storage; s2, generating a statistical row vector by combining row gradient aggregation with a historical attenuation factor; s3, carrying out block distributed statistics in the column direction and synchronously generating column vectors across equipment; s4, a low-rank matrix is constructed through outer products of row and column vectors, and estimation precision is improved through noise suppression; s5, performing dynamic sparse sampling, performing initial high-density focusing, and stabilizing the sampling rate of a key layer; s6, the sampling points execute time sequence attenuation updating, and asynchronous calculation is carried out to improve the resource utilization rate; s7, performing Gaussian kernel smoothing on a neighborhood value compensation coverage gap in an unsampled region; s8, fusing low-rank estimation and sparse data, and balancing global precision by self-adaptive weight; the method has the beneficial effects of reducing video memory occupation, and improving distributed computing efficiency and balance training precision and speed.

The invention relates to a quantum chemistry Hamiltonian rapid generation method and device based on a stability sub-table, and the method comprises the steps: obtaining a single electron integral item, a double electron integral item and a nuclear energy item of a target Fermi Hamiltonian, and storing the nuclear energy item; dividing the equal double-electron integrals into the same calculation object class, and deleting double-electron integral items corresponding to the double-electron integrals of which the absolute values are smaller than a threshold value; converting the single-electron integral item and the double-electron integral item into a Bose Hamiltonian expression form; filling the stable sub-form table based on all Fouli items to obtain a Fouli check matrix, and recording a coefficient vector; the same row vectors in the Follicle check matrix are merged, the Follicle coefficients corresponding to the merged row vectors are added, and the row vectors with the Follicle coefficients being 0 are deleted; and storing the Follicle check matrix and the coefficient vector. According to the technical scheme, the efficiency of converting the Fermi Hamiltonian into the Bose Hamiltonian can be improved.

The invention relates to the technical field of server aging testing, in particular to a server aging testing method and device, electronic equipment and a storage medium. Then convolution processing is carried out on each aging data queue in a sliding mode through a difference operator, the obtained multiple first convolution queues are segmented one by one, queue segments obtained through segmentation are normalized, and multiple queue segments are obtained; performing dot product calculation on each queue segment, each same-time-segment queue segment and each same-part queue segment to obtain a plurality of dot product results, and adding representative values determined according to the plurality of dot product results into a test analysis matrix; and finally, according to the plurality of row vectors of the test analysis matrix, determining the abnormal probability of the plurality of tested parts of the server. According to the method, the state of the detected component is determined based on the cross point product and probability estimation, so that the problem component can be found early.

The invention provides an SAR (Synthetic Aperture Radar) anti-interference evaluation method and device based on maximum variance combined with layered entropy weight, and relates to the technical field of radar anti-interference. Comprising the following steps: constructing an evaluation index matrix according to a plurality of target SAR echoes; dividing the plurality of evaluation indexes into two layers of indexes by using a Pearson correlation analysis algorithm according to the evaluation index matrix and a preset threshold value; according to the plurality of evaluation indexes, the random number and the normalized evaluation index matrix, constructing a maximum variance evaluation optimization model which takes a maximized weighted evaluation matrix and a utility value matrix as objective functions and takes a difference degree weight row vector as a parameter to be solved; according to the maximum variance evaluation optimization model and the evaluation index matrix, determining a maximum difference evaluation matrix by using an interior point method; and according to the maximum difference evaluation matrix, the evaluation index matrix and the two layers of indexes, determining a plurality of evaluation values corresponding to the plurality of anti-interference methods by using a layered entropy weight method. In this way, the accuracy of anti-interference evaluation of various anti-interference methods is high.

The invention provides a game behavior anomaly detection method based on spatio-temporal feature fusion, which comprises the following steps of: visual coding of behavior data: collecting user operation time sequence data, and performing hue mapping coding according to event types, respectively mapping the basic operation event, the core interaction event, the numerical value change event and the non-operation event into differentiated colors, wherein the numerical value event adopts a gradient color to represent the change intensity; dividing the time sequence data into row vectors based on a preset time unit, generating a two-dimensional behavior map through multi-row stacking, transversely representing time continuity, longitudinally reflecting behavior cycle regularity, and forming a visual data carrier containing spatial-temporal characteristics; and two-channel detection model construction: constructing a double-layer architecture of a supervised detection channel and an enhanced detection channel, wherein the supervised detection channel adopts a deep convolutional neural network to extract map spatial features and is connected with an interpretable classifier to output a behavior compliance probability.

The invention provides a motion trail prediction method and system based on a human body space-time interaction relationship, and belongs to the technical field of motion trail prediction. Comprising the steps of performing three-dimensional human body posture estimation and two-dimensional point estimation on a court where players are located based on a sports competition video, and extracting motion posture data and track coordinates of the players; extracting a posture interaction relationship based on the motion posture data, and calculating a posture relationship adjacency matrix between players; extracting spatial features and time domain features in the attitude relation adjacency matrix by using a three-dimensional convolutional neural network, and solving the interaction strength between players to obtain a player interaction relation adjacency matrix; the player interaction relation adjacency matrix is converted into a group of row vectors to serve as time sequence interaction relation vectors of all players, and the row vectors and the trajectory coordinates of all the players are fused into input vectors of a trajectory prediction model; and inputting the player input vector and the time sequence interaction relationship vector into a trajectory prediction model for prediction to obtain a future motion trajectory coordinate considering the interaction relationship between players.

The invention discloses an infrared image vertical stripe removing method based on dynamic adjustment of vertical stripe distribution. The method comprises the following steps: (1) a high-dimensional feature vector matrix extraction step, and (2) a vector parameter extraction step: inputting the high-dimensional feature vector matrix and each frame of to-be-processed infrared image into a vertical line removing parameter generation network together, and generating two column vector parameters and two row vector parameters; 3) an improved traditional vertical stripe removing algorithm processing step which comprises the steps of performing Gaussian filtering on the infrared image to be processed in the row direction, calculating vertical stripe noise column by column, judging whether the vertical stripe noise is real vertical stripe noise column by column, obtaining a vertical stripe noise image of the image to be processed, and then performing subtraction to obtain an infrared image after vertical stripe removal; according to the method, a proper loss function and a proper training process are designed for the vertical stripe distribution information extraction network and the vertical stripe removal parameter generation network, so that the model achieves good performance, the distribution condition of vertical stripes can be effectively analyzed, and robust vertical stripe removal parameters are generated.

The invention discloses a method for constructing a defect detection model, and a defect detection method and device. The method for constructing the defect detection model comprises the following steps: acquiring a target defect image of a display panel; according to the brightness data of each pixel in the target defect image, constructing a row vector sample and a column vector sample, and respectively carrying out classification labeling on the row vector sample and the column vector sample; constructing a classification model for defect detection, wherein the classification model comprises a support vector machine classification model in a row direction and a support vector machine classification model in a column direction; respectively acquiring a first classification hyperplane of the support vector machine classification model in the row direction and a second classification hyperplane of the support vector machine classification model in the column direction; and generating a defect detection model storing the first classification hyperplane parameter and the second classification hyperplane parameter. According to the method for constructing the defect detection model and the defect detection method and device provided by the invention, local defects in the display panel can be detected.

Systems and methods embedding a frequency domain-based watermark in tabular data are provided. In some methods, a covariance matrix is computed for the tabular data. Row vectors of the tabular data are projected onto two orthonormal vectors e1, e2 of the covariance matrix. A signal space for the projected row vectors is determined. A frequency f* with a lowest power from the signal space is located. A watermark is embedded in the tabular data based on the frequency f*.

The invention relates to the technical field of lake and reservoir water bloom prediction, in particular to an intelligent lake and reservoir water bloom prediction system and method based on a DTW algorithm. Comprising a data acquisition module, a matrix construction module, a weight distribution module and a display alarm module. Concentration values of chlorophyll a at different positions of a reservoir are collected through a data collection module, the concentration values are arranged according to a time sequence and divided into two parts, a matrix building module is used for calculating the similarity between a comparison sequence Xc and a training data set Xt, a prediction matrix is built with a time step length n as a row vector and a sub-sequence with the similarity m as a column vector, and a prediction result is obtained. According to the model, only chlorophyll a is adopted as monitoring data, the requirement for original data is greatly reduced, the calculation amount of the model is reduced through the DTW algorithm, and the accuracy and robustness of a prediction result are guaranteed by dynamically adjusting the weight between different similarities.

The invention provides a water supply network water leakage abnormity identification method based on NMF dimension reduction, and relates to the technical field of water supply network abnormity identification, and the method comprises the steps: embedding hydrophone sample data containing expert labeling information into a sample weight matrix as a row vector, and constructing a sample dimension loss item; hydrophone sample data containing industry feature priori are mapped to the feature basis matrix to serve as rows, and feature dimension loss items are constructed; combining the two loss items to construct a total loss function and solving the total loss function to obtain an optimal sample weight matrix; and performing clustering analysis based on the optimal sample weight matrix to obtain a clustering affiliation result of the hydrophone sample data so as to judge whether the water supply network to be identified has water leakage abnormity or not. According to the method, the double-constraint NMF framework fusing the expert knowledge of the sample dimension and the industry priori knowledge of the feature dimension is constructed, the accuracy and interpretability of hydrophone feature dimension reduction are considered, and the problems that an existing method is poor in interpretability, less in priori knowledge utilization and the like are solved.

The invention discloses an intelligent table data escape and knowledge base construction method based on a large language model. The method comprises the steps that multiple document formats are imported; identifying the existence of the table, analyzing the row-column structure of the table, positioning the position of each cell, and extracting data in the table; the big language model understands the data significance of each cell according to the context and the data content in the table; performing escape on the original data; merging and storing the original data, logically partitioning the data, and vectorizing the data after escaping; table data in a knowledge base is obtained through a large language model knowledge base system, and a query request of a user is quickly responded through a vector index of the knowledge base. According to the method, the data in the table is intelligently converted by adopting a large language model semantic escape technology, and the data is embedded into the knowledge base through data merging operation, so that a user can conveniently query and utilize the data, and the problem of insufficient table data processing precision in the prior art is solved.

The invention discloses a fault positioning method of a nuclear power instrument control system, a related device and a storage medium, and the method comprises the steps: constructing a correlation matrix model of the nuclear power instrument control system, and constructing an application correlation matrix based on the correlation matrix model and the configuration information of a target control station; obtaining the current operation state information of the nuclear power instrument control system, and constructing a current health state matrix by using the current state of each test point in the current operation state information; matching the current health state matrix with an application correlation matrix; if the current health state matrix is only matched with one row vector, taking the equipment corresponding to the row vector and the information of the fault mode of the equipment as a current fault positioning result; if the current health state matrix is matched with the plurality of row vectors, calculating a priority index corresponding to the equipment based on the failure probability and maintenance difficulty of the equipment corresponding to each row vector; and according to the priority indexes corresponding to the equipment, outputting the equipment corresponding to each row vector, the information of the fault mode of the equipment and the corresponding priority indexes.

The invention relates to the technical field of robot control, and discloses a robot control method and system based on multi-modal fusion, and the method comprises the steps: obtaining multi-modal sensing data, extracting features, and carrying out the clustering, thereby obtaining an associated information cluster sequence; a cluster center vector is extracted and quantized, and a quantitative evaluation matrix is obtained; calculating a covariance between matrix row vectors, and if the covariance is greater than zero, marking the covariance as a positive correlation coordination component to obtain a coordination component set; grouping the coordination component set, and if the number of the components in the group is less than a threshold value, merging adjacent groups; selecting a central group and predicting a change trend, and optimizing robot parameters after evaluating a risk value according to the change trend; and optimizing the updated parameters according to the multi-modal sensing data and a preset updating amplitude, and if the difference between the parameters before and after updating exceeds a threshold value, reducing the updating amplitude, and readjusting the parameters to obtain a final decision model. According to the method, the problem of low robot group cooperation efficiency in a complex scene can be solved.

The embodiment of the invention provides a light spot feature calculation method, device and equipment for a dynamic image, and is applied to the technical field of optical image processing, and the method comprises the steps: carrying out the summation of each row of a non-zero pixel matrix, obtaining a column vector, multiplying the column vector by an index row vector, dividing the product sum by the sum of non-zero pixels, and obtaining the light spot feature of the non-zero pixel matrix. Obtaining a y coordinate value; summing each column of the non-zero pixel matrix to obtain a row vector, multiplying the row vector by the index column vector, and dividing the product sum by the sum of the non-zero pixels to obtain an x coordinate value; according to the x coordinate value and the y coordinate value, x-direction centroid coordinates and y-direction centroid coordinates of the light spot are obtained; calculating the variance of the x direction and the y direction according to the x coordinate value and the y coordinate value, and calculating the light spot roundness and the light spot diameter based on a least square method according to the variance of the x direction and the y direction; and calculating a beam waist radius and a beam divergence angle based on the light spot diameter. In this way, the real features of the light spots in the dynamic image can be accurately calculated.

The invention discloses a method for improving photoetching performance of a dipole light source. The method comprises the following steps of: finding out a selected area with an OPC (Optical Proximity Correction) segment coupling effect in a layout graph, selecting a plurality of adjacent selected OPC segments in the selected area, and combining the selected OPC segments to form a segment group; the row number and the column number of the Jacobian matrix are equal to the number of the selected OPC segments, and each element of a row vector of the Jacobian matrix is a derivative of the light intensity of one selected OPC segment corresponding to the row vector to the position coordinate of each selected OPC segment; each element of a column vector of the Jacobian matrix is a derivative of the light intensity of each selected OPC segment to the position coordinates of one selected OPC segment corresponding to the column vector. A Jacobian matrix is added into a cost function, then MB-OPC correction is carried out, and a light source in an OPC model subjected to MB-OPC correction is a dipole light source. According to the invention, the OPC convergence speed of an area with a strong OPC segmented coupling effect can be accelerated, and the contrast ratio can be improved.

The invention discloses a calculation method and device for realizing Fourier transform by a parallel circuit. The method comprises the following steps: acquiring an input signal sequence of an input signal; based on a preset degree of parallelism, performing serial-parallel conversion on the input signal sequence by using a parallel circuit to obtain a signal transformation matrix; obtaining an original Fourier transform matrix according to the input signal sequence, and transforming the original Fourier transform matrix into a parallel Fourier transform matrix based on the degree of parallelism; and performing vector inner product operation on the row vectors of the parallel Fourier transformation matrix and the signal transformation matrix in sequence to obtain a Fourier transformation result of the input signal sequence. During calculation, it is not required that the length of an input signal sequence must be N power of 2 or 4, and it is not required that the input signal sequence must be uniformly sampled, but a Fourier transform result can be obtained by sequentially carrying out vector inner product operation on row vectors of the parallel Fourier transform matrix and the signal transform matrix, so that the calculation efficiency is improved, and the calculation cost is reduced. And the waste of computing resources is avoided.

The present invention discloses a method for minimal-delay fine-grained parallel Givens rotation of ultra-large-scale matrix QR decomposition, relating to the technical field of signal and information processing. The method comprises: an outer loop and an inner loop, wherein the outer loop is responsible for calculating a row vector of a Q matrix and an R matrix, and the inner loop is responsible for the Givens rotation of the two row vectors; the outer loop is decomposed into multiple loop blocks and a multi-level PE unit array is used to realize parallel operation of the outer loop slices, while the parallel calculation of the inner loop slices is realized by a step-by-step decreasing element update module inside the PE unit; the collaborative design of the inner and outer loop slices significantly reduces the idle time of hardware resources and optimizes the generation and transmission of key data, so that the system starts calculation with minimal delay; the method improves the utilization efficiency of hardware accelerators and is suitable for ultra-large-scale matrix decomposition tasks in the fields of signal processing, image processing, and scientific computing.

The invention discloses an LDPC (Low Density Parity Check) code check matrix construction method, which is used for processing a low earth orbit satellite ground foundation zone, and comprises the following steps of: A1, determining degree distribution of a basis matrix under a high code rate by adopting an extrinsic information transfer graph EXIT, and realizing construction of the basis matrix under the constraint of optimal degree distribution through extrinsic information transfer P-EXIT based on a basic model graph; a2, based on the code length and the code rate of FEC coding of the DVB-S2 protocol, determining the size and the expansion factor of a basis matrix corresponding to each code rate; a3, determining a finite field and a generator based on an expansion factor, and constructing a linearly independent shift value row vector and cycle coefficient table; a4, selecting row vectors from the cyclic coefficient table to construct a cyclic shift matrix, and performing short ring identification and elimination on the cyclic shift matrix; and A5, performing matrix hashing on the basis of the basis matrix and the cyclic shift matrix to obtain an LDPC check matrix.

The invention belongs to the technical field of defect detection, and particularly relates to a defect detection method and system in a filter assembly laminating process based on image processing, so as to solve the technical problems that an existing detection method model is relatively weak in micro defect sensing capability and relatively low in detection accuracy. The detection method comprises the following steps: S1, carrying out block attention calculation on an image block sequence; s2, calculating an adjustment coefficient and adjusting the updating process of the row normalization factor by using the adjustment coefficient; s3, weighting corresponding row vectors in the initial context feature matrix by using the adjustment coefficient corresponding to each query block to obtain a weighted context feature matrix; and S4, outputting a mask pattern representing the position of the defect on the surface of the filter assembly. According to the detection method provided by the invention, secondary weighting and feature fusion are carried out on the initial context features, so that the accuracy of defect detection of the filter assembly is improved.

The invention provides a matrix multiplication calculation method and device based on a three-dimensional accelerated calculation architecture, and the method comprises the steps: segmenting a quantized target matrix into bit slice matrixes along a bit dimension, and carrying out the clustering of the bit slice matrixes along a row vector dimension, and obtaining a clustering center matrix and a class residual matrix; combining repeated operation between different columns of the clustering center matrix to obtain an intermediate activation vector; multiplying the enumeration matrix by the intermediate activation vector to reconstruct a calculation result of the clustering center; multiplying the similar residual matrix by the input activation, and adding with the calculation result of the corresponding clustering center to obtain the calculation result of the similar residual matrix; and obtaining a multiplication operation result of the target matrix and the input activation based on calculation results of all the clustering centers and the class residual matrix. According to the method, through joint optimization of three dimensions of bit, row and column, the redundancy calculation is minimized, the number of necessary operations is effectively reduced, and the complexity and calculation overhead of large-scale matrix multiplication are reduced.

The present application is applicable to the technical field of integrated circuits and neural networks. Provided are a matrix multiplication apparatus and method based on a systolic array, and an electronic device. The matrix multiplication apparatus based on a systolic array comprises: a systolic array formed by means of arranging processing elements, wherein two adjacent processing elements in each column are connected by a shift, each processing element has transmission channels in a row direction, a column direction and a diagonal direction, the row direction is used for inputting a row vector of a first matrix, and the column direction and the diagonal direction are used for inputting a column vector of a second matrix; a plurality of first control signal channels, which are used for inputting a first control signal for controlling the operating state of each shift; and a plurality of second control signal channels, which are used for inputting a second control signal for controlling each column vector of the second matrix to be input in the column direction or the diagonal direction. Therefore, a hardware structure taking into consideration a sparse matrix multiplication operation and a dense matrix multiplication operation is achieved, and the calculation complexity of matrix multiplication is reduced.

The present invention discloses a DOA estimation method for a single-bit sensor array, which relates to the field of signal processing technology, including obtaining an original signal, performing single-bit quantization on the original signal to obtain a quantization result, and calculating the covariance matrix of the quantization result; vectorizing the covariance matrix to generate a column vector; performing DFT operation on the column vector to obtain spectrum information; obtaining the peak value and the peak position of the spectrum information through a findpeaks function, processing to obtain the arrival angle position, and finding the left and right boundary positions of the arrival angle position to determine the arrival angle interval; using DTFT to perform frequency refinement on the arrival angle interval, and performing spectrum peak search on the refined arrival angle interval, calculating and obtaining the final angle estimation. The present invention has low computational complexity, and the estimated angle is more accurate than the same type of algorithms, and can be extended to other nonlinear array structures.

A device for optical matrix-vector multiplication is provided. A x → = y → , encompassing: n * m optically coupled cells (i,j) (330, 331-399) representing the elements of an (n×m) matrix A, an input vector x → , whose components x i (221,222,223) are represented by n waveguides which are coupled to the rows of matrix A, forming an output vector y → whose components y j (231,232,233) are represented by m waveguides coupled to the columns of matrix A, wherein each cell (i,j) (330, 331-399) comprises an optical amplifier (400) configured to apply a weighting factor a ij to implement ∈[1;∞[.

The invention discloses a super-large scale matrix QR decomposition minimum delay fine granularity parallel Givens rotation method, and relates to the technical field of signal and information processing, the method comprises an outer layer cycle and an inner layer cycle, the outer layer cycle is responsible for calculation of a row vector of a Q matrix and a row vector of an R matrix, and the inner layer cycle is responsible for Givens rotation of the two row vectors; an outer layer cycle is decomposed into a plurality of cycle blocks, parallel operation of an outer layer cycle piece is achieved through a multi-stage PE unit array, and meanwhile parallel calculation of an inner layer cycle piece is achieved in a PE unit through a step-by-step decreasing element updating module; the idle time of hardware resources is remarkably shortened through the collaborative design of the inner circulation piece and the outer circulation piece, generation and transmission of key data are optimized, and the system starts calculation with the minimum delay; the method improves the utilization efficiency of the hardware accelerator, and is suitable for super-large-scale matrix decomposition tasks in the fields of signal processing, image processing, scientific calculation and the like.

The application discloses a new correlation image recognition and machine learning method based on fuzzy monotony, and belongs to the technical field of artificial intelligence of pattern recognition and machine learning; the method is defined as FMMCA, which evaluates local fuzzy monotone correlation by comparing row vectors and column vectors of an image matrix pair by pair, then the local correlations are weighted and accumulated, and finally the global fuzzy monotone correlation between images is obtained. The application directly uses fuzzy monotone correlation analysis to replace classical correlation analysis for multi-view research, so that the problems existing in classical correlation analysis do not exist, and the fuzzy monotone method feature does not need to be measured statically by distance, but can be measured dynamically by interval change, so that the influence of some noise is offset, the performance is improved, a new fuzzy monotone machine learning method is formed, the optimization of a focus loss function is not needed, the parameters are few, the robustness is good, and the computing power is small.